Upon obtaining approval from the Ethics Committee of The Affiliated Hospital of Xuzhou Medical University (ethics approval number : XYFY2024-KL254) and informed consent from the patients, we conducted a retrospective analysis of clinical and imaging data from 18 patients diagnosed with MCI and treated with craniectomy combined with rapid internal decompression (CCRID) between March 2022 and March 2024 (Fig. 1). The surgical protocol was in accordance with the guidelines for surgical treatment of MCI [5].

Inclusion criteria were 1) preoperative computed tomography angiography (CTA) indicating MCA occlusion; 2) CT confirms acute malignant cerebral infarction with involvement of at least two-thirds of the territory supplied by the MCA; 3) comatose state; 4) midline shift greater than 0.5 cm and/or the presence of significant intracranial hypertension and/or progressive neurological deterioration despite aggressive medical treatment; 5) intraoperatively, brain tissue elevated above the bone window by 1-2 cm with poor pulsation; and 6) informed consent obtained from the patient’s family regarding the surgical procedure and associated risks.

Exclusion criteria were 1) fixed and dilated pupils bilaterally; 2) unstable vital signs; 3) absence of spontaneous respiration or other signs of brain death; and 4) uncorrectable coagulopathy or other severe complications that preclude surgical intervention.

For 18 cases with CCRID (Table 1), the data extracted from medical records included patient demographics (gender and age), preoperative and postoperative National Institutes of Health Stroke scale (NIHSS) and Glasgow coma scale (GCS) scores, details of EVT, and postoperative average ICP. Imaging reports provided information on the affected cerebral hemisphere, preoperative and postoperative midline shift distances, and hemorrhage in the infarcted area post-surgery. Intraoperative records documented the distance of brain tissue protruding beyond the edge of the bone window after DC, cerebral hemisphere resection time, and total surgical time. All patients were followed up for 3 months postoperatively. Meanwhile, we statistically compared preoperative and postoperative NIHSS, GCS, midline shift distances, ICP, perioperative complications, and 3-month postoperative modified Rankin scale (mRS) scores of these patients with those of 24 patients who underwent DC combined with partial temporal/frontal pole resection (DCPTR) at our hospital from March 2019 to March 2022.

Categorical data were analyzed using chi-square tests, continuous data with t-tests, and non-parametric data with ranksum tests. Comparisons between groups were conducted using the chi-square test for categorical variables and either unpaired t-test or Mann-Whitney U test for continuous variables depending on the distribution of the data. Statistical significance was set at p<0.05.

Under general anesthesia, the patient was positioned supine with the head securely fixed. A large, question mark-shaped scalp incision was made, and the scalp flap was reflected to expose the skull. Burr holes were drilled, and a craniotome was used to create a 12-15 cm bone flap, which was carefully removed. The dura mater was then incised and reflected to expose the swollen brain (Fig. 2A).

Partial temporal lobe resection is divided into a dominant hemisphere and non-dominant hemisphere. The range of temporal pole resection in the dominant hemisphere includes 4 cm in the anterior part of the temporal lobe, starting from the superior temporal gyrus, deep to the parahippocampal gyrus, and bottom to the tentorial edge of the cerebellum. The decompression range of the non-dominant hemisphere includes 6 cm in the anterior temporal lobe, starting from the superior temporal gyrus, deep to the parahippocampal gyrus, and bottom to the tentorial edge of the cerebellum.

According to the plane of the upper margin of the bone window, the extent of brain tissue excision was determined, and the arachnoid membrane was delineated using bipolar electrocoagulation (Fig. 2B). With the cooperation of two experienced neurosurgeons, the brain tissue was excised at the upper edge of the flat bone window using nerve strippers, bipolar coagulation, suction devices, scissors, and other instruments. Tiny blood vessels and tough connective tissues were carefully severed after electrocauterization. The entire brain tissue removal process took approximately not exceed 10 minutes, followed by meticulous hemostasis (Fig. 2C and Supplementary Video 1). After thorough irrigation with normal saline and hydrogen peroxide, ensuring no active bleeding, an artificial dural membrane was applied. The bone flap was not replaced. An ICP monitoring tube and an epidural drainage tube were placed. Finally, the scalp was closed in a routine manner, and the patient was transferred to the intensive care unit for monitoring and treatment.

From March 2022 to March 2024, 18 patients diagnosed with MCI underwent CCRID. The cohort had an average age of 62.89 years, with a male-to-female ratio of 11 : 7. Thirteen patients had left hemisphere infarctions, while five had right hemisphere infarctions. The mean preoperative NIHSS score was 20.22, and the mean preoperative GCS score was 8.39. Detailed patient characteristics and intraoperative findings are presented in Table 1. Postoperative outcomes, including ICP, midline shift at 1 and 7 days post-CCRID, and NIHSS and GCS scores at 2 weeks, were systematically recorded. The mean postoperative ICP was 4 mmHg. There was a significant reduction in midline shift postoperatively, with a mean shift of 0.23 cm at 1 day and a further reduction at 7 days. Postoperative NIHSS and GCS scores improved to means of 17.22 and 10.11, respectively. No cases of intracranial infection were observed. There was one case of postoperative intracranial hemorrhage (ICH) and 2 mortalities within the follow-up period. The mean mRS score at 3 months was 4.45. These data are summarized in Tables 2 and 3.

The patient presented with altered consciousness and right-sided hemiparesis for 5 hours before admission. CTA revealed a core infarct volume of approximately 37 mL and a mismatch volume of about 216 mL (Fig. 3A). The patient promptly underwent arterial thrombectomy and stent placement in the M2 segment of the MCA, performed by two neurologists and interventional radiologists (Fig. 3B). Postoperatively, tirofiban was administered at a continuous infusion rate of 8 mL/h. Despite these interventions, the patient remained in a state of severe coma 2 days later, with bilateral pupil diameters enlarged to approximately 4 mm. Follow-up CT imaging indicated an increase in the infarcted area with a midline shift of approximately 10 mm (Fig. 3C-E).

An immediate DC was performed. A large bone flap measuring approximately 12×13 cm was removed, revealing bulging infarcted brain tissue protruding about 2.5 cm beyond the bone edge (Fig. 3F). Decompression of the bone window was conducted, and the procedure was completed in approximately 5.7 minutes (Fig. 3G and H). Hemostasis was meticulously achieved, and the surgical site was covered with a gelatin sponge, hemostatic gauze, and artificial dura mater before closing the cranium (Fig. 3I). An ICP monitor was placed, and the patient was transferred to the ICU for continued care.

On the first postoperative day, the average ICP was approximately 3 mmHg. Follow-up CT showed no hemorrhage in the infarct area and an 8 mm reduction in the midline shift (Fig. 3J-L). By the fifth postoperative day, the midline shift had completely resolved (Fig. 3M and N). Over the subsequent three months, the patient did not experience perfusion hemorrhage, intracranial infection, or severe brain edema.

For comparison, data from 24 patients who underwent DCPTR were analyzed. The mean age in this group was 60.17 years, with a male-to-female ratio of 11 : 13 and a mean preoperative NIHSS score of 22.92. Postoperative outcomes, including NIHSS, GCS, midline shift, ICP, perioperative complications, and 3-month mRS scores, were compared between the two groups. Statistical analysis revealed no significant difference in NIHSS, GCS, midline shift, or ICP between the two groups, although the incidence of ICH and mortality rate was higher in the DCPTR group. The mean mRS score at 3 months was slightly higher for the DCPTR group but not statistically significant. These comparisons are detailed in Table 3.

MCI, or malignant cerebral infarction, results from the occlusion of major cerebral arteries, causing extensive brain tissue damage, severe cerebral edema, and elevated ICP [14]. This condition presents with significant neurological deficits and has a high mortality rate, often exceeding 80% with conservative management alone [10]. According to current guidelines, the indication for EVT primarily focuses on patients with small to moderate-sized cerebral infarctions [11]. However, some trials have demonstrated that EVT in patients with acute ischemic stroke and large infarct cores results in better functional outcomes compared to medical therapy alone, although it carries a higher risk of ICH [4,15,21]. Surgical decompression not only mitigates the life-threatening risks associated with brain herniation following extensive cerebral infarction but also enhances penumbra perfusion with early intervention, thereby significantly improving the patient’s functional prognosis [7,23]. At present, the guidelines have not strongly recommended whether to further perform internal decompression after craniectomy because of the lack of long-term and large-scale effectiveness, and the method of internal decompression has not yet been unified [22]. In our clinical process, we found that DC did not bring effective decompression in some cases, and there have also been some literatures recommending internal decompression [16,27]. Our study demonstrates that CCRID may be an effective surgical technique for patients with MCI necessitating internal decompression.

The reason why we chose a flat bone window for decompression is because : a) the brain tissue higher than the bone window during surgery is mostly necrotic brain tissue, and there is little bleeding during the internal decompression surgery; b) the height of brain tissue above the bone window is generally 1-2 cm. Removing this part of brain tissue will not damage the important structure thalamus or open the lateral ventricle, reducing the possibility of intracranial infection after ventricular opening; and c) we removed brain tissue through a flat bone window to alleviate overall brain pressure. As for the removal of the anterior temporal lobe, it is to reduce brainstem compression caused by local temporal lobe infarction. Since uncinate gyrus of temporal lobe is closest to the brainstem, the herniation of the uncinate gyrus tends to occur easily, leading to respiratory and cardiac arrest in patients. Based on DC, we further performed anterior temporal lobectomy. For the safety of patients, we will strive to achieve a comprehensive reduction in both overall and local pressure. Our findings indicate that CCRID could significantly reduce ICP and midline shift, thereby improving clinical outcomes [20,25]. This technique may present a safer and more efficient alternative to other surgical methods. Compared to traditional DC, CCRID not only involves the removal of a standard large bone flap but also incorporates rapid internal decompression by excising brain tissue through a flat bone window. However, our technique is primarily applicable when the infarcted area encompasses the region that requires resection, which can be estimated preoperatively using CT imaging. The CCRID technique offers several advantages over DCPTR. Our study found that the mean operation time for CCRID was shorter than that for DCPTR (2.82 vs. 3.23 hours), attributable to the rapid cerebral hemisphere resection. This reduction in surgical time may decrease the risk of perioperative complications and improve overall patient outcomes and reduced healthcare costs. Additionally, the incidence of intracranial infections was lower in the CCIRID group, further supporting the potential benefits of this technique. Unfortunately, patient 12 with CCRID passed away more than a month after discharge. Although the CT scan at discharge indicated that the midline shift had not fully resolved, suggesting inadequate decompression, our follow-up revealed that the patient ultimately succumbed to worsening pulmonary infection. This was due to the family’s inability to afford the necessary long-term care and treatment.

This distinction sets our method apart from other internal decompression techniques. Of course, this is only a preliminary and rough standard, and further standard specifications need to be further improved in our subsequent research. Additionally, the timing of decompression remains a critical factor that warrants further investigation.

The goal of surgical treatment for MCI is to reduce ICP, increase cerebral perfusion, and save brain tissue within the penumbra. However, craniotomy for MCI is associated with several common complications that must be meticulously managed to optimize patient outcomes. ICH is a significant concern, as bleeding within the brain or surgical site poses a substantial threat to recovery [12]. Patient 4 developed intracerebral hemorrhage after CCRID, perhaps due to excessive decompression and reperfusion injury to blood vessels. Fortunately, the bleeding did not persist. Infection, including surgical site infections and intracranial infections such as meningitis or abscess formation, is another critical complication [12]. Additionally, complications such as hydrocephalus [19,24], postoperative seizures [3,6], subdural hematomas, or delayed cerebral ischemia associated with the bone flap are relatively common [18]. Notably, none of our patients experienced malignant cerebral edema. Unfortunately, two patients in this study succumbed to severe pulmonary infections within 3 months of post-surgery. Despite the small sample size, the 3-month survival rate for patients receiving CCRID reached 90%, indicating that our study warrants further investigation.

Despite the promising findings, our study has several limitations. Firstly, the sample size was relatively small, and the study was conducted at a single institution, which may limit the generalizability of the results. Secondly, the retrospective nature of the study may introduce selection bias. Prospective, multicenter studies with larger sample sizes are necessary to validate our findings and establish standardized protocols for CCRID. Another limitation is the lack of long-term follow-up data beyond 3 months. While our study provides valuable insights into the short-term outcomes of CCRID, future research should focus on evaluating the long-term efficacy and safety of this technique, including functional outcomes and quality of life measures. We aim to address these limitations in subsequent research.